The Spatial and Temporal Variation of Measured Urban PM10 and PM2.5 in the Helsinki Metropolitan Area

We have studied particulate matter (PM) concentrations,PM₁₀ and PM_2.5, measured in an urban air qualitymonitoring network in the Helsinki Metropolitan Area during1997–1999. The data includes PM₁₀ concentrationsmeasured at five locations (two urban traffic, one suburbantraffic, one urban background and one regional backgroundsite) and PM_2.5 concentrations measured at twolocations (urban traffic and urban background sites). Theconcentrations of PM₁₀ show a clear diurnal variation,as well as a spatial variation within the area. Bycontrast, both the spatial and temporal variation of thePM_2.5 concentrations was moderate. We have analysedthe evolution of urban PM concentrations in terms of therelevant meteorological parameters in the course of oneselected peak pollution episode during 21–31 March, 1998.The meteorological variables considered included wind speedand direction, ambient temperature, precipitation, relativehumidity, atmospheric pressure at the ground level,atmospheric stability and mixing height. The elevated PMconcentrations during the 1998 March episode were clearlyrelated to conditions of high atmospheric pressure,relatively low ambient temperatures and low wind speeds inpredominantly stable atmospheric conditions. The resultsprovide indirect evidence indicating that the PM₁₀concentrations originate mainly from local vehiculartraffic (direct emissions and resuspension), while thePM_2.5 concentrations are mostly of regionally andlong-range transported origin.     

Characteristics of air pollution in Beijing during sand-dust storm periods

In the Beijing area, March and April have the highest frequency of sand-dust weather. Floating dust, blowing sand, and dust storms, primarily from Mongolia, account for 71%, 20%, and 9% of sand-dust weather, respectively. Ambient air monitoring and analysis of recent meteorological data from Beijing sand-dust storm periods revealed that PM₁₀ mass concentrations during dust storm events remained at 1500 μg m⁻³, which is five to ten times higher than during non-dust storm periods, for fourteen hours on both April 6 and 25, 2000. During the same period, the concentrations in urban areas were comparable to those in suburban areas, while the concentrations of gaseous pollutants, such as SO₂, NO _x , NO₂, and O₃, remained at low levels, owing to strong winds. Furthermore, during sand-dust storm periods, aerosols were created that consisted not only of many coarse particles, but also of a large quantity of fine particles. The PM_2.5 concentration was approximately 230 μg m⁻³, accounting for 28% of the total PM₁₀ mass concentration. Crustal elements accounted for 60–70% of the chemical composition of PM_2.5, and sulfate and nitrate for much less, unlike the chemical composition of PM_2.5 on pollution days, which was primarily composed of sulfates, nitrates, and organic material. Although the very large particle specific surface area provided by dust storms would normally be conducive to heterogeneous reactions, the conversion rate from SO₂ to SO₄ ²⁻ was very low, because the relative humidity, less than 30%, was not high enough.     

The Relative Influences of Primary and Secondary Particulates to Urban Air Quality in the United Kingdom

This study uses a combination of data from U.K. monitoringstations and from modelling undertaken with the U.K.Meteorological Office's NAME Model to investigate therelative influences of primary and secondary particulateson total PM₁₀ levels at sites in the United Kingdom. Co-located PM₁₀ and sulphate aerosol measurementsindicate that sulphate has a disproportionately largeinfluence on the variation of PM₁₀ levels incomparison to its contribution to their total mass.Comparisons of measured PM₁₀ at urban centre, roadsideand rural sites suggest that local primary sources havevery little influence on daily mean levels. NAME has beenused to model both primary particles and sulphate aerosolfrom sources across the whole of Europe. The discrepanciesbetween modelled and observed PM₁₀ suggest that coarseparticles, such as windblown dust and resuspended roaddust,may comprise a very large, if not dominant, proportion ofobserved PM₁₀ levels. The apparently minor role ofprimary particles (especially locally-sourced ones) raisesa number of issues regarding the suitability of current U.K.and European legislation to addressing the particle problem.     

Comparison of Four Machine Learning Methods for Predicting PM10 Concentrations in Helsinki, Finland

Machine learning methods can offer a practicalalternative to deterministic and statistical methods forpredicting air pollution concentrations. However, for agiven data set, it is often not clear beforehand whichmachine learning method will yield the best predictionperformance. This study compares the variable selection andprediction performance of four machine-learning methods ofdifferent complexity: logistic regression, decision tree,multivariate adaptive regression splines and neuralnetwork. The methods are applied to the task of predictingthe exceedance of the European PM₁₀ daily averageobjective of 50 g m^-3 for a station in Helsinki,Finland. Our study shows that some predictors were selectedby all models but that the different models also pickeddifferent variables. The performance of three of the fourmethods investigated was very similar, however, performanceof the decision tree method was significantly inferior.Performance was sensitive to the learning sample size andtime period used.     

ROADWAY-2: A Model for Pollutant Dispersion near Highways

The formulations and evaluation of ROADWAY-2, a near-highway pollutant dispersion model, are described. This model incorporates vehicle wake parameterizations derived from canopy flow theory and wind tunnel measurements. The atmospheric velocity and turbulence fields are adjusted to account for velocity-deficit and turbulence production in vehicle wakes. A turbulent kinetic energy closure model of the atmospheric boundary layer is used to derive the mean velocity, temperature, and turbulence profiles from input meteorological data. ROADWAY-2 has been evaluated using SF₆ tracer data from General Motors Sulfate Dispersion Experiment. The model evaluationresults are presented and discussed.     

Identification,Abundance and Origin of Aliphatic Hydrocarbons in the Fine Atmospheric Particulate Matter of Athens,Greece

The organic chemical composition of the fine fraction of atmospheric particulate matter in Athens has been studied, in order to establish emission sources. The results of the analyses of the aliphatic fraction indicate that all samples contain n-alkanes ranging from C₁₄ to C₃₂, with C₂₅, C₂₆, C₂₇ and C₂₉ being the more abundant congeners. Fossil fuels biomarkers such as extended tricyclic terpanes (hopanes, steranes) and isoprenoid hydrocarbons (pristane, phytane) were observed in our samples on a daily basis. Source reconciliation was conducted using molecular diagnostic ratios (such as the carbon preference index – CPI). The mean CPI value (1.84) indicates the mixed origin of the Athenian fine particles. The notable presence of an unresolved complex mixture or “hump” of hydrocarbons in our gas chromatograms is indicative of petrogenic hydrocarbon inputs. An approximate measure of this kind of contamination is the ratio of the concentrations of unresolved components to the resolved n-alkanes and other major compounds (U:R). The high U:R value of 25.25 further confirmed the major contribution of fossil fuels. Yet, the percent contribution of leaf wax n-alkanes (25.15%) indicated the parallel contribution of biogenic sources. This work supports the conclusion that vehicular emissions were the major source of aliphatic organic compounds with a smaller contribution of biogenic n-alkanes during the study period in Athens.